Friday, July 24, 2015

John McCain

I have never understood why people love, or even tolerate, John McCain.  That's not to say that Trump's thuggish attacks are justified.  But on the merits, why think John McCain is good at anything?

I have a suspicion that a group of Dem "dirty tricks" folks posed as Repubs, and tried to get McCain the nomination as a joke. And they are still laughing that it worked.

There is really no other plausible explanation. You can just hear the meeting, after lots of beers:

"Okay, seriously, what idiot can we try to get them to nominate?"

"Jeb Bush?" "No, Bush name is toxic now in 2007. No way they'll go for it."

"I've got it! No, really, I'VE GOT IT! John.....McCain!" (Everybody in room breaks up laughing). "No....the Repubs are stupid, but they aren't THAT stupid."

"Seriously, it's genius. He calls himself a 'war hero,' and nobody in that country club bunch would have the balls to challenge him on his actual policy views. And all the other candidates will split the vote of the conservative voters. It's perfect."

(And so they agreed, and McCain was the nominee). I bet it happened just that way.

Thursday, July 23, 2015

Why Mr. Overwater Invented the Internet....

...so we can watch a guy inside a giant water balloon that explodes in slow motion.  The balloon, not the guy.

Wednesday, July 22, 2015

Follow up on the Nerdy Rant

When I posted my nerdy rant about the importance of going for efficiency, I mentioned my research with my friend Rodolfo Cermeno on GARCH in panels. However in the published version, Badi made us cut all the simulation exercises.

Just this morning though, I found a draft of the paper with the simulations contained.

We first compared OLS to our MLE panel garch estimator where the conditional variance was correctly specified and we found that,


"when comparing the OLS and MLE estimators (for the mean equation), we find that the MLE outperforms the OLS estimator in terms of bias, precision and mean squared error. In every sample, the MLE estimator has a MSE smaller than the OLS estimator by at least a factor of 4 when ρ =.25 and at least a factor of 5 when ρ = .5"


ρ is the arch(1) coefficient.

We then compared OLS to a mis-specified panel garch estimator and found that for the conditional mean coefficients, the MSE was lower with the misspecified garch estimator by at least a factor of 2.5.

Those are pretty big MSE improvements!


When the variance reductions are big and the bias induced in the conditional mean from a mis-specified conditional variance model is small, OLS with asymptotically correct standard errors is kind of a dumb way to conduct research.


Tuesday, July 21, 2015

KPC Guide to the 2016 Prez Race: Elect-ageddon!

Rather than give actual information about candidates, which would violate our basic tenet ("Don't vote, it just encourages the bastards!"), I want to propose a fieldwork version of the "Grand Game."

The Grand Game is where you read a linked article, or piece in the media, and compete to find the most absurd quote, or compete to ask the most devastating question.  It's...well...Grand!

So, readers, go out into the world and observe the behavior (mating and otherwise) of the wild creatures called candidates, and report back.  My email is mcmunger (at) gmail [dot] com.  Just send in your entries.

And here is the judging criteria, from the incomparable H.L. Mencken.  Find the words and actions of candidates that best embody the wisdom of this quote:

The state — or, to make matters more concrete, the government — consists of a gang of men exactly like you and me. They have, taking one with another, no special talent for the business of government; they have only a talent for getting and holding office. Their principal device to that end is to search out groups who pant and pine for something they can’t get, and to promise to give it to them. Nine times out of ten that promise is worth nothing. The tenth time it is made good by looting ‘A’ to satisfy ‘B’. In other words, government is a broker in pillage, and every election is a sort of advanced auction on stolen goods. (Notes on Democracy, 1926).
 It was "men" in his time, of course, but Ms. Pelosi and Ms. Clinton are better men--in the above sense--than most men in Mencken's time.

Of course, Mencken also said that "Democracy....is the worship of jackals by jack asses."  That'll get 'er done, too.

Go forth, readers, and report back!  I'll post everything you find for the next 15 months, until "Elect-ageddon 2016!"

(Lagniappe:  Hilariously, Sheldon had the same "voter guide" thought, and used the same Mencken quote.  It's a MOVEMENT, people!)

Monday, July 20, 2015

Phone Performance Art

You may recall this incident.  It was obviously pretty extreme.

But my own experience just now cancelling my Wall Street Journal subscription wasn't much different.  I called to cancel, and of course it's a bad sign that you have to call.  It would be easier and faster, and cheaper for them, actually, if you could cancel on the web site.  So the only possible explanation is that they think that they can harass and intimidate you into not cancelling.  The people on the phone are hired thugs.

Knowing this, I was ready. Still, it took a little more than 21 minutes to get the job done.  I won't go through the entire conversation, but it went like this (several times):
________________________________________
WSJ Thug:  "Can you tell my why you want to cancel your subscription?"

Me:  "Absolutely not."

WSJT:  ... (clearly has been trained that silence is power)

Me:  ...(has known for a long time silence is power)

WSJT:  "But we want to make sure we provide the best service we can to our customers.  We want to know how we could do better."

Me:  "That makes sense.  Tell you what, give me your fax number.  I have a contract [rattle piece of paper near phone, audibly] for consumer service consulting.  $175 per hour, four hour minimum, payable in advance.  As soon as I get your check or money order, I'll be happy to answer your questions for up to 4 hours."

WSJT: ...(not sure what to do, because this is not on his script).

Me:  "Or you can let me talk to your supervisor, right now.  Just forward this call to your supervisor."

WSJT: ....

Me:  "Or, you can cancel my subscription."

WSJT:  "I can't cancel your subscription.  I need to ask why you want to cancel."

Me:  "And I'm happy to answer that question.  But I told you I need a check for $700 first."

WSJT:  "Why would we pay you?"

Me:  "Why would I provide your free customer service consulting?  Do YOU work for free, pumpkin?"

WSJT: ...


[We went through this exchange, almost verbatim, just repeating things word for word, three times. Finally...]

WSJT:  "Okay, I'll cancel the subscription.  But what if I offer you the lowest rate, $16.00 per month?"

Me:  "Are you going to cancel my subscription?"

WSJT:  "That's up to you, sir.  I'm offering you the lowest rate!"

Me:  "No, it's not up to me.  I asked you to cancel the subscription 20 minutes ago, and you have been harrassing me and refusing to do what I want.  So, it's clearly up to you, not me."

WSJT: ...

Me:  "Tell  you what.  I'll offer you a discount.  Just $150 per hour, three hour minimum.  So you can send me a check for $450, and I'll answer your questions.  Otherwise, please put me through to your supervisor."

WSJT: "Well, the reason we aren't going to send you a check isn't the cost.  We don't want to do that at all."

Me:  "And that's exactly how I feel about your newspaper.  Why would a discount change my mind?  Would you please cancel my subscription?"

__________________________________________

He finally did.  Remarkable.  I wish I had thought to record it.  The aggressive thuggery is really out of place at a newspaper that presents itself as professional.  I will certainly never subscribe to any WSJ products again.

Friday, July 17, 2015

Political Science: The Best & Brightest

Look at who is the latest political science major:



Yes people, that's  Anders Breivik, newly matriculated at the University of Oslo to major in......

.....Political Science!!!

The future of the discipline is in good hands.

*****Bonus Snark********

Do you know what Norway gave him for killing 77 people?

21 years.

That's a little over  3 months per victim.

Wednesday, July 15, 2015

Who cares about getting an asymptotically correct standard error for a possibly massively inefficient coefficient estimate?

Well sure,  Angrist and Plishke and their legion of devotees, but that barn has already been burned by my main man, Edward Leamer (see the section entitled "White-washing).

But now Cameron & Miller take up the cudgel to argue for OLS plus "cluster-robust" standard errors.


Let's review:

1. Under Heteroskedasticity, OLS is unbiased but inefficient (has a variance larger than the minimum variance unbiased estimator) and normal OLS standard errors are biased.

2. The robust standard errors crowd ignores the first issue to focus on the second, but do not produce such a great answer even to the problem of the OLS standard errors, because their robust standard errors have only an asymptotic justification (i.e. the only property that can be shown for them is consistency).

3. Leamer has argued that the first problem is the more important problem. Since some forms of heteroskedasticity make OLS extremely inefficient, we need to find a better estimator. These are often called FGLS (feasibly generalized least squares) estimators, where the researcher estimates a model of the conditional variance along with the model for the conditional mean.

4. Here is where the robust standard errors folks raise the bogeyman of bias in their argument against  looking for a better estimator.

Here's the money quote from Cameron & Miller:

"One way to control for clustered errors in a linear regression model is to additionally specify a model for the within-cluster error correlation, consistently estimate the parameters of this error correlation model, and then estimate the original model by feasible generalized least squares (FGLS) rather than ordinary least squares (OLS). ... If all goes well this provides valid statistical inference, as well as estimates of the parameters of the original regression model that are more efficient than OLS. However, these desirable properties hold only under the very strong assumption that the model for within-cluster error correlation is correctly specified."

Look people, we have two enemies when we try to get a point estimate of an unknown parameter, variance and bias.

Suppose you don't have the exactly correct functional form of the conditional variance but you do FGLS anyway. Say you create 10% bias by doing so. It is still the case that the reduction in the variance may well be sufficient to accept that increased bias and use the mis-specified FGLS estimator instead of the least squares estimator.

In my own research work on GARCH models with Rodolfo Cermeno, we show that even incorrectly specifying the conditional variance often produces coefficient estimates superior on mean squared error grounds to OLS.

This happens because of the extreme inefficiency of OLS in the face of some forms of heteroskedasticity and because small mis-specifications of the conditional variance model do not seem to lead to large biases in the estimates of the conditional mean parameters.

Furthermore, you actually can perform some statistical tests to see how well your chosen model of the conditional variance is working. Simply use the standardized FGLS errors and test them for general forms of heteroskedasticity and see what happens. If we fail to reject the null of no heteroskedasticity at say the .25 level, we are pretty confident in our functional form.

Finally, since all we know about robust standard errors is that they are consistent, in a finite sample the OLS + robust errors approach can give us a very inefficient parameter estimate and a biased standard error for that parameter estimate.

This problem is even greater in the case of clustered residuals because the requirement becomes that the number of clusters goes to infinity, not just the number of observations!

Double finally, let me note (and Camerer & Miller do a good job of explaining this), FGLS and clustered standard errors are not mutually exclusive. You can do both. My recent paper with Dan Hicks and Weici Yuan is an example.

Mrs. Angus has just informed me that this piece should be titled, "The Nerdiest Rant Ever".

But people, variance is just as big a problem as bias and consistency alone is a weak reed on which to base your estimation strategy.

Mrs. Angus has just informed me that I've just proven her point with the previous sentence.

Monday, July 13, 2015

Fair is Fair, But It Depends on Where



Fair Is Not Fair Everywhere 
Marie Schäfer, Daniel Haun & Michael Tomasello
Psychological Science, forthcoming

Abstract: Distributing the spoils of a joint enterprise on the basis of work contribution or relative productivity seems natural to the modern Western mind. But such notions of merit-based distributive justice may be culturally constructed norms that vary with the social and economic structure of a group. In the present research, we showed that children from three different cultures have very different ideas about distributive justice. Whereas children from a modern Western society distributed the spoils of a joint enterprise precisely in proportion to productivity, children from a gerontocratic pastoralist society in Africa did not take merit into account at all. Children from a partially hunter-gatherer, egalitarian African culture distributed the spoils more equally than did the other two cultures, with merit playing only a limited role. This pattern of results suggests that some basic notions of distributive justice are not universal intuitions of the human species but rather culturally constructed behavioral norms.

Saturday, July 11, 2015

GOP: The More Candidates, The Scarier

From my friend David Frist over at AHI, a piece on "Real Clear Politics."

Excerpt:

An important factor in lengthening the candidate roster and also making a quick winnowing unlikely is a long decline in political discipline among conservatives, who for many decades have dominated the Republican electorate. One notable result of this indiscipline is trouble judging who is most worth backing in a presidential race – the proliferation of fuzzy thinking about who is most likely to win a general election, remain true to conservative principles, and deliver for conservatives as president. This situation results from at least two causes. One is conservatives' and libertarians' ambivalent attitude toward power and therefore toward practical, as distinct from merely expressive, politics. The other is their long record of frustration with presidential power and federal authority.

Friday, July 10, 2015

Love Gov: The Series

Here's the first installment of the quite funny, yet disturbingly creepy, "Love Gov" series from the Independent Institute:




You can find the rest of the them here.  Worth watching them all. 
Premise:  clever. 
Execution:  so accurate that it's a little uncomfortable.